EdgeProof Lab Strategy Robustness Tester
- 实用工具
-
Karem Neili
I build professional MT5 tools for traders of indices, gold and FX: Expert Advisors, indicators and utilities focused on risk control, multi-timeframe analysis and prop-firm account protection.
My principles:
• No repainting: signals are based on closed candles only. - 版本: 1.0
- 激活: 10
A good backtest is not proof. It can come from luck, from a few outsized trades, from costs that are too optimistic, or from parameters fitted to the past. EdgeProof Lab takes the closed trades of any strategy and puts them through eight stress tests, then gives one clear robustness score from 0 to 100 with a grade from A to F.
THE TESTS
- Monte Carlo drawdown: 2,000 resampled trade sequences. Shows the 95th percentile drawdown compared with the backtest, and the share of paths that end in profit.
- Trade order: the same trades in random order. Shows whether the backtest drawdown was simply a lucky sequence.
- Remove the best trades: drops the best 5% of trades (adjustable). If the profit factor collapses, the edge depends on a few lucky trades.
- Extra costs: adds a cost per lot to every trade (spread, commission, slippage) and shows how much of the profit survives.
- Missed trades: randomly skips 15% of trades (adjustable), as happens live with requotes, VPS downtime or filters.
- Stability over time: how many quarters were profitable.
- Statistical significance: t-statistic of the average trade, with a warning when there are too few trades.
- Optimization plateau: read an optimization export from the Strategy Tester. If the neighbouring parameter values lose most of the profit, the best result is an isolated peak, a sign of curve fitting.
Each test shows PASS, WARN or FAIL with the numbers behind it, so you know exactly what to fix.
TRADES TO ANALYZE
- Your account history, filtered by magic number, symbol and start date. Run your EA on a demo account or use the history of a live account.
- Or a CSV file (close_time, symbol, type, volume, net, magic) in the common Files folder, for trades from any other source.
- The starting balance is detected from your deposits, or entered by you.
OUTPUT
- A clear panel on the chart with the score, a Monte Carlo fan chart and the test table.
- Save HTML report: a clean report you can open in any browser and share.
- Export baseline: saves the trades as a baseline file for EdgeProof Monitor (live drift monitoring).
HOW TO USE
1. Attach EdgeProof Lab to any chart.
2. Set the magic number of the strategy (or -1 for all trades) and, optionally, symbol and start date.
3. Optional: in the Strategy Tester, right-click the optimization results, choose Export to XML, save the file in the common Files folder (Terminal\Common\Files\EdgeProof\) and enter its name, e.g. EdgeProof\opt.xml.
4. Read the score and the warnings. Click Re-run for new random paths.
GOOD TO KNOW
- EdgeProof Lab never opens, changes or closes trades.
- At least 10 closed trades are needed; 100 or more give meaningful results.
- Drawdowns are measured in money as a percent of the starting balance, so reordering trades does not change their weight.
- If no trades are found, the panel shows a sample strategy marked SAMPLE DATA. The Strategy Tester always shows the sample strategy, because the tester has no trade history to read.
- The score is a guide based on past trades. It is not a guarantee of future results.
